IBM and Microsoft Just Gave Enterprises 6,000 Pre-Built Agents

AI Agents

IBM and Microsoft Just Gave Enterprises 6,000 Pre-Built Agents

What happened: On Tuesday, IBM and Microsoft unveiled Imagine Enterprise, a joint platform for building, running, and governing AI agents inside industry-specific workflows. IBM Consulting contributes more than 6,000 pre-built agents built on its Consulting Advantage platform; Microsoft supplies the foundation — Foundry, Copilot Studio, Fabric, Entra, Purview, and Defender. Early targets are financial services, insurance, and retail, with engineering support to move customers from pilots to production.

Why it matters: This is the clearest signal yet that the enterprise AI market is shifting from “experiment with a chatbot” to “stand up an agentic workforce.” The pitch isn’t a faster assistant — it’s an agentic business, as IBM Consulting’s Mohamad Ali put it. The real story is the bundle: pre-built agents plus governance and security baked in, aimed squarely at buyers who can’t afford to assemble agents from scratch. SEI, a fintech managing roughly $2 trillion in assets, is already exploring the stack.

What’s next: Watch whether “6,000 agents” translates into recurring revenue or stays stuck in exploratory spend. IBM’s consulting revenue was flat last quarter; Microsoft’s Azure grew 48%. Neither disclosed pricing or adoption targets — the honest question is whether enterprises can turn this into measurable ROI, not agent counts.

Source: Microsoft

Adobe is adding native support for the Model Context Protocol (MCP) across Photoshop, Illustrator, Premiere, and After Effects — letting outside agents operate directly on native project files while keeping layers and edits reversible. It’s the biggest shift to Creative Cloud since 2012, and it reframes the creative tool as a platform agents plug into, not just an app humans drive. Source: Creative Boom

Meta’s Muse AI agent is rolling out to Apple Watch and Wear OS, letting users place voice calls, dictate, or text the assistant to book reservations, manage email, and order groceries. Muse hit 10 million U.S. downloads in under 40 days — a pace only 11 other apps have matched — signaling that consumer AI is moving from novelty chat to a genuinely ambient, always-on assistant. Source: TechCrunch

AI News

Using AlphaProof Code, a multi-agent framework, DeepMind mathematically proved the security properties of the actual Signal protocol code — the first time a widely deployed cryptographic protocol has been formally verified at this depth. The system generated over six million lines of Lean and flagged just 1% of the codebase for human review, compressing what once took years into four weeks. Source: Google DeepMind

AI training-data company Mercor raised $750 million at a $20 billion valuation, doubling in a year, with Nvidia, Salesforce, and Zoom among investors. The company, which pays a network of over five million experts to generate training data and evaluations, claims it scaled from $1M to $3B in run-rate revenue in 27 months — proof that the “labeling” layer of the AI stack is becoming a core industry, not a commodity. Source: Unite.AI

Cisco is integrating Anthropic’s Mythos model into AI Defense to generate multi-step, system-specific attacks that probe enterprise AI agents before adversaries do. Access starts with a limited set of enterprises — a reminder that as agents get more autonomy, “red-teaming” them is becoming a standard part of the security stack. Source: eWeek